Re: c/4844: Code generation bug for hppa64

Re: c/4844: Code generation bug for hppa64

[ross.alexander]

The following reply was made to PR c/4844; it has been noted by GNATS.

From: ross.alexander@uk.neceur.com
To: gcc-gnats@gcc.gnu.org
Cc: law@cygnus.com
Subject: Re: c/4844: Code generation bug for hppa64
Date: Mon, 26 Nov 2001 13:18:40 +0000

 To all concerned,
 
 After some investigation I have tracked the problem down to
 expr.c.  The statement
 
      b.neg =  !b.neg;
 
 Somewhere this generates a tree EQ_EXPR.  In expand_expr
 the do_store_flag (line 7896) fails and the code falls through to
 TRUTH_ANDIF_EXPR which as the line
 
      if (target)
           emit_clr_insn(target);
 
 Unfortunately as the target is also the RHS on the expr it clears
 the value before it gets a chance to all emit_0_1_insn.
 
 I don't really have a good enough understanding of what is going
 on to do anything more about this.
 
 Any help would be most appriciated.
 
 Many thanks
 
 Ross
 
 ---------------------------------------------------------------------------------
 
 Ross Alexander                           "He knows no more about his
 MIS - NEC Europe Limited            destiny than a tea leaf knows
 Work ph: +44 20 8752 3394         the history of East India Company"
 
 
                                                                                                                      
                     gcc-gnats@gcc                                                                                    
                     .gnu.org             To:     Ross Alexander <ross.alexander@uk.neceur.com>                       
                                          cc:                                                                         
                     09/11/01             Subject:     Re: c/4844: Code generation bug for hppa64                     
                     17:56                                                                                            
                     Please                                                                                           
                     respond to                                                                                       
                     gcc-gnats;                                                                                       
                     Please                                                                                           
                     respond to                                                                                       
                     nobody                                                                                           
                                                                                                                      
                                                                                                                      
 
 
 
 
 Thank you very much for your problem report.
 It has the internal identification `c/4844'.
 The individual assigned to look at your
 report is: unassigned.
 
 >Category:       c
 >Responsible:    unassigned
 >Synopsis:       Code generation bug for hppa64
 >Arrival-Date:   Fri Nov 09 09:56:01 PST 2001